Goal: reduce the roundtrip request/response time of Retool RPC resource without causing crashes or other issues.
Context: our backend server hosting Retool RPC runs on Fastify on Railway.app with no auto-sleep.
———
Hi all!
Does anyone know whether lowering the polling interval for Retool RPC down to 10 or 5ms would cause any issues on the Retool side?
Currently, the minimum polling interval is 100ms, which often results in 300-500ms response times (I assume the minimum is 200ms for roundtrip) for very simple queries (e.g. return the string “hello world”).
The processing time on both the retool side and our servers is < 10ms each, so I believe the polling time makes it slow.
My main concern is whether this would be more likely to cause server errors on either side (since that already happens).
Any thoughts? Do the JS and Python implementations differ at all in this regard?
Lowering polling intervals is generally not recommended. Can you clarify more about your environment?
Are you an on-prem or cloud user? When you mention "multiple RPC instances," are you referring to multiple RPC agents? If so, it's important to note that the agents themselves control their own polling and throughput. Queries remain in a queue until an agent polls and fetches them.
Using multiple agents is totally fine—that’s actually something we do ourselves as well.